Relocation Allowances and Damage Awards
David L. Beattie*
The object of damages in personal injury
suits is to place the plaintiff in the position
he or she would have been in but for the
injury, so far as that is possible by the pay-
ment of a monetary sum. If the plaintiff is
forced to relocate in order to obtain medical
or other necessary services, the calculation of
an appropriate sum is more complicated be-
cause the cost of living may be different in
the new place of residence. In this note, the
author examines two methods of calculating
adjustments to damages awards to reflect dif-
ferences in the cost of living. The compen-
sating cost approach examines differences in
the prices of the various items upon which
the plaintiff would spend his or her income
(including taxes), and the award is adjusted
to reflect these differences. The compensating
income approach compares the incomes of
persons performing similar functions in the
two areas, and adjusts the award accordingly.
The author argues that although the com-
pensating income method is relatively simple
to calculate, it does not always compensate
changes in the cost of living accurately. The
compensating cost approach is more difficult
to calculate, but satisfactory statistical series
do exist.
Dans une poursuite pour pr6judice corporel,
les dommages accord~s visent ft placer le dc-
mandeur dans la situation qu’il ou elle cOt
occup6e n’eflt W du pr6judice, cn tenant
compte des possibilit6s offertes par une
compensation mon~taire. Si le demandeur
doit etre relocalis6 pour b6n~ficier de soins
m6dicaux ou autres services ncessaires, des
differences dans le coOt de la vie au nouveau
lieu de residence pourraient compliquer le
calcul de l’indemnit6. Dans cette note, I’au-
teur traite de deux m6thodes de calcul des
rajustements des indemnit6s en consequence.
La technique de < compensation des coflts >
rajuste l’indemnit6 A partir des differences de
prix pour divers postes de d6pense du de-
mandeur (incluant les taxes et l’imp~t). La
technique de ((compensation du revenu > se
base sur les differences de revenu pour des
personnes occupant des fonctions semblables
dans chaque lieu afin de rajuster l’indemnit6.
Uauteur soutient que, quoiqu’elle se calcule
plus simplement, la m6thode de compensa-
tion du revenu ne rflte pas toujours les
changements dans le coOt de la vie de fagon
exacte. Les calculs sont plus compliqu~s sous
la m6thode de compensation des coOts, mais
il existe des bulletins statistiques ad~quats
pour les faciliter.
*Department of Economics, University of Regina. I would like to thank my colleagues within
the Department and my legal associates for providing constructive comments on earlier ver-
sions of this paper.
19871
Introduction
NOTES
This paper examines the unique problems which arise in damage cases
where the injured party, as a direct result of an accident, is forced to change
location (place of residence). While perhaps reflecting the lack of readily
available statistics, or simply the infrequency of such cases, the potential
importance of locational cost differences as an element of damages appears
to have attracted little attention. This situation sharply contrasts with tra-
ditional concerns over estimating price changes over time, where damage
awards by utilizing real rates of discount provide allowances for intertem-
poral cost differences.’
The rationale for relocation allowances and their implications for dam-
age awards are presented in Section I. In Sections II through IV, two alter-
native frameworks (a compensating cost approach and a compensating
income approach) are developed to estimate spatial differences in the cost
of living. Both frameworks also attempt to address the constraints under
which legal processes operate. The methodologies to the extent possible must
not only generate accurate estimates, but also must provide for time-efficient
estimation given the current state and availability of data bases. The con-
cluding section of the paper presents a comparative evaluation of the two
approaches.
To assist in the development of the general arguments, as well as to
provide illustrative estimates of the various spatial costs, the following hy-
pothetical case will be considered.
An Ottawa teacher has been rendered a quadriplegic as a result of an injury.
The Court, after consideration of the evidence on pecuniary and non-pecuniary
losses (including contingencies), has arrived at a judgement as to their present
values, that is, the lump-sum awards for each head of damage. One remaining
potential head of damage has yet to be evaluated. The injured individual will
be required to move to Vancouver so as to access the necessary medical facilities
and support services.
At the outset, two points require emphasis. First, the magnitude of the
resulting cost estimates are not in themselves significant. Given that damage
estimates must be as case-specific as possible, this illustration is only in-
tended to serve as a guide or case example. Application of the methodologies
and data bases to cases involving different occupations or locations will
necessarily produce different relocation allowances. Second, in the interests
of brevity, the examination of pecuniary damages has been deliberately
restricted to losses involving prospective income. However, the spatial ar-
guments, and the significance of such allowances, also apply to other pe-
cuniary losses in personal injury cases (for example, health care costs), as
‘See S. Waddams, The Law of Damages (Toronto: Canada Law Book, 1983) at 248-50.
McGILL LAW JOURNAL
[Vol. 33
well as to pecuniary losses in fatal accident cases (where dependants have
been forced to relocate).
I. Relocation Allowances: Rationale
As outlined by Bruce,2 when personal injury damages solely involve
lost income, the standard objective is to estimate a single self-extinguishing
lump-sum payment which, along with interest earnings, will replace the
stream of lost income (the earnings the individual would have received had
the damage not occurred). If the intent of such awards is to provide to the
degree possible an equivalent standard of living, then spatial differences in
the cost of living cannot be ignored. Damage payments, by utilizing statistics
and estimates which are location-specific, presume the injured party will
continue to live in the location from which the data were drawn. In terms
of the above example, the estimate of the growth rate in the base-period
income loss would employ salary information on Ottawa teachers. Similarly,
the estimate of the discount (real interest) rate would employ Ottawa-specific
price data. The resulting award assumes the individual would continue to
live in Ottawa, purchasing goods and services at Ontario prices. However,
if a move to Vancouver is now required, and if British Columbia prices
exceed Ontario prices for a comparable bundle of commodities, then the
lump-sum award would not provide an equivalent standard of living. In
this case, the purchasing power of the payment for lost wages would only
allow a smaller basket of goods and services to be purchased through time.3
The need to assess locational price differences is not only stressed by
economists, but also by jurists. As Cooper-Stephenson and Saunders state:
Again, increased costs of living incurred because of the necessary relocation
of the plaintiff will be allowable. A required move from a rural to an urban
location close to hospital and medical facilities might give rise to substantially
increased costs … . In short, all post-trial home and living costs incurred by
the plaintiff’s accident are properly assessable under this head.4
2C. Bruce, “The Calculation of Foregone Lifetime Earnings: Three Decisions of the Supreme
Court of Canada” (1979) 5:2 Can. Pub. Pol. 155 at 156; Assessment of Personal Injury Damages
(Toronto: Butterworths, 1985) c. 2.
3The use of Canada average statistics would not solve the problem of spatial price differences.
Such statistics would not account for cost of living differences when a change of residence
involved at least one location either above or below the national average. Furthermore, failure
to use location-specific and occupation-specific data would place in question the lost income
estimate. Possible distortions are introduced if the underlying wage and price statistics fail to
reflect the particulars of the damage case under consideration.
4K. Cooper-Stephenson & I. Saunders, Personal Injury Damages in Canada (Toronto: Cars-
well, 1981) at 316-17.
1987]
NOTES
II. Compensating Cost Versus Compensating Income
Before detailed presentations are made of the compensating cost and
compensating income frameworks, it is useful to outline briefly the two
methodologies. The compensating cost approach directly examines the var-
ious factors which give rise to cost of living differences. Estimates of the
cost differences are then used to derive the relocation allowance. In contrast,
the compensating income approach employs a proxy measure of the cost
differences. If regional wage levels reflect the underlying costs of living, then
income differentials for comparable employees in each location provide an
indirect estimate of the relocation allowance.
III. The Compensating Cost Approach
In terms of design, not estimation, the compensating cost approach is
relatively straightforward. First, the major factors contributing to spatial
differences in the cost of living are identified. In this paper, they are grouped
under three headings:5 annual consumption expenditures, shelter/housing
costs, and taxation, transfers, and public goods. Second, estimates are then
made of any cost differences. Finally, employing the spatial estimates, the
award for lost income is adjusted to offset cost increases arising from the
change of residence.
1. Annual Consumption Expenditures
This component of the relocation allowance examines variations in the
prices of goods and services, specifically, annual expenditures on consump-
tion items. Two relationships must be estimated. First, a cost estimate is
required of the difference in consumer prices across the two locations. Sec-
ond, the relative importance of the cost estimate must be determined. Since
expenditures on goods and services only account for a fraction of total family
expenditures, a calculation of the proportion of income directed toward
consumption is required.
To examine the extent to which prices of goods and services vary by
location, use is made of the Statistics Canada series entitled “inter-city retail
price differentials.”‘ 6 These statistics provide place-to-place comparisons of
the absolute differences in retail prices. Excluding shelter (expenditures to
5The cost groupings were determined by the structure and coverage of existing data bases.
6It may be useful to note that the Statistics Canada data on consumer price indices for major
Canadian cities are not helpful when evaluating spatial price differences. These statistics provide
city-specific measures of price changes through time. They allow comparisons of the rates of
price change across cities, not comparisons of their relative price levels. See generally Statistics
Canada, Consumer Prices and Price Indexes (Ottawa: Minister of Supply & Services) (62-0 10;
Quarterly) app. I.
REVUE DE DROIT DE McGILL
[Vol. 33
own or rent accommodations), 7 this expenditure-weighted series utilizes
price data on approximately eighty percent of the items surveyed by the
consumer price index. The surveys are constructed such that similar qual-
ities of goods and services are compared at a common moment in time.8
Inter-city price differentials have been published since 1975 and are reported
in index form, that is, each city’s value is expressed relative to an all-city
average equal to 100.9
Table 1 reports values of two composite indices for the Ottawa/Van-
couver example. Percentage differences are also provided over the 1975-
1985 time period. The estimate of cost of living differences should not be
dominated by possible aberrations in any given year’s statistics. Just as the
other data underlying damage awards (the estimates of wage rate changes,
rates of price change, and interest rates) are to reflect their expected values
during the period of lost income, spatial estimates also are to be repre-
sentative of the expected differences in price levels. In the present case,
employing the eleven years of available data, Vancouver prices on average
7While household operation and furnishings are included in the surveys, the omission of
shelter costs (on average accounting for 27.6% of total expenditures) is significant. Separate
treatment will be given to the 19.0% share associated with rent, mortgage, and property tax
payments. See below, Part III, Sections 2 and 3.
8Fixed basket indices contain weighting biases. If consumers substitute away from com-
modities which become relatively more expensive, then indices which employ base period
(Laspeyres) weights will tend to overstate the impact of price changes. An opposite result will
hold if current period (Paasche) weights are employed. Recent studies, however, suggest that
the biases will be minimized if the underlying weighting structures are frequently updated. The
upward bias in the consumer price series has been estimated to be in the order of 0.1% per
year. See: P. Genereux, “Impact of the Choice of Formulae on the Canadian Consumer Price
Index” in W. Diewert & C. Montmarquette, eds, Price Level Measurement (Ottawa: Minister
of Supply & Services, 1983) 489; Statistics Canada, The Consumer Price Index Reference Paper,
Updating Based on 1982 Expenditures (Ottawa: Minister of Supply & Services, 1985) (62-553)
at 95-108.
9The indices are constructed for major urban centers throughout Canada; they do not provide
for urban-rural comparisons. For a listing of the surveyed cities, as well as the commodity
groupings covered, see notes a-c of Table 1. Inter-city price differentials were available in two
prior years (1969 and 1971); Winnipeg, however, was employed as the base equal to 100.
The Conference Board of Canada also surveys living cost differentials for selected Canadian
cities. Apart from differences in objective and methodology (in particular, the Board’s treatment
of shelter costs/capital appreciation and their omission of transfer payments and public goods),
the Conference Board estimates provide point-in-time cost comparisons for two stylized house-
holds. See J. Frank, ProvincialDifferences: A Challenge to Compensation and Relocation Policies
(Ottawa: Conference Board of Canada, 1981); M. Daniel, “Comparative Living Costs: Who’s
Ahead?” (1982) 9:1 Can. Bus. Rev. 44; L. Threlfall, “How Have Comparative Living Costs
Changed Since 1981?” (1983) 10:3 Can. Bus. Rev. 29.
1987]
NOTES
exceeded Ottawa prices by 5.0% per year.10 Therefore, to provide for sim-
ilarity in living standards, the consumption share of the annual income
losses (the proportion of income spent on consumption items) requires an
upward adjustment in the order of 5.0% per year.
At this point, it is useful to identify two requirements of the con-
sumption share estimate. First, the estimate should capture the intertem-
poral nature of consumer decisions. Consumption expenditures not only
refer to monies spent on current year consumption, but also include ex-
penditures which will serve to maintain or supplement consumption levels
in the future, for example, annual savings for retirement. Second, the es-
timate should be case-specific, reflecting the plaintifis expenditure patterns
as determined by such variables as income and family size.
Representative data on the apportionment of income by expenditure
category are also published by Statistics Canada.” Numerous cross-
tabulations are provided in these statistics –
expenditure distributions by
family size, age of head, income level, and form of shelter accommodation.
Therefore, in order to provide an illustrative consumption estimate, addi-
tional detail must be given to the hypothetical accident case.
Assume the teacher held an Ontario Class 2-3 designation (one year of teacher
training and a four-year university degree).’ 2 Also, at the date of accident
(1985), the individual was thirty-nine years old, married with two dependent
children (minors). Finally, a home was being purchased and ten years remained
before expected mortgage payout.
Table 2 reports the Statistics Canada data specific to the above case.
These statistics are based upon the most recent (1982) survey of family
‘0Comparative data on rates of price change suggest that this estimate of price level differences
has remained relatively stable through time. For the 1975-1985 time period, the annual rates
of change in the consumer price index for Ottawa and Vancouver were very similar, equalling
8.14% and 8.07% respectively. If the time period is expanded to include 1965-1985, the per-
centage change statistics continue to exhibit only marginal differences: Ottawa (6.69%) and
Vancouver (6.66%). See also M. Denny & M. Fuss, “Regional Price Indexes: The Canadian
Practice and Some Potential Extensions” in Diewart & Montmarquette, supra, note 8, 783.
While this estimate represents a simple average of past cost differences, it does not preclude
the use of more sophisticated estimation approaches, for example, distributed lag models (Box-
Jenkins techniques).
“Statistics Canada, Family Expenditure in Canada (1982) (Ottawa: Minister of Supply &
Services, 1984) (62-555). Notwithstanding questions of availability, the consumption share
estimate does not preclude the use of information drawn directly from the plaintiff’s accounting
records.
‘Qualifications Evaluation Council of Ontario, Teachers’ Qualifications Evaluation Pro-
gramine 3 (1975) at 11; Ontario Secondary School Teachers’ Federation, Secondary School
Teacher Certification (1979) at 19. Also see infra, note 40.
McGILL LAW JOURNAL
[Vol. 33
TABLE 1
INTER-CITY RETAIL PRICE DIFFERENTIALS,
1975-1985 (COMBINED CITY AVERAGEa = 100)
Year
Composite Index
Composite Index
Value
Value
C.P.I. Weightsb
Survey Weightsc
Percentage
Difference
Ottawa Base
Ottawa Vancouver Ottawa Vancouver
C.RI.
Wts.
6.4%
6.3%
5.2%
5.4%
4.5%
2.2%
2.9%
6.2%
5.3%
6.5%
6.0%
Survey
Wts.
5.7%
5.9%
4.7%
4.7%
4.4%
1.9%
2.5%
5.9%
4.9%
6.1%
5.4%
98.2
98.2
98.1
97.4
99.3
99.9
99.8
98.3
98.6
98.7
97.9
97.7
98.1
97.8
97.1
99.3
99.8
99.7
98.4
98.7
98.7
97.9
104.0
104.3
102.9
102.3
103.8
102.0
102.6
104.5
103.9
105.1
103.8
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
Annual
Ave.:
aStatistics Canada bases the combined city average for each grouping of goods and services
upon the following surveyed cities: St. John’s, Charlottetown, Halifax, Saint John, Montreal,
Ottawa, Toronto, Winnipeg, Regina, Edmonton, and Vancouver.
103.8
104.0
102.7
102.0
103.7
101.8
102.3
104.1
103.4
104.7
103.2
103.6
5.2%
4.8%
98.6
103.3
98.5
bThe index values for the various commodity groupings were combined into a composite
index by employing the representative expenditure weights underlying the consumer price
index. In 1985, these Canada average weights were: food (20.0%), shelter (27.696), household
operation and furnishings (10.6%), clothing (8.4%), transportation (15.8%), health and personal
care (4.0%), recreation, reading and education (8.3%), and tobacco and alcohol (5.5%).
CDue to non-comparabilities, the inter-city retail price statistics are not based upon a complete
set of commodities representative of total consumer expenditures. To examine the impact of
this partial sampling, the “illustrative” survey weights reported by Statistics Canada have been
employed to construct a second composite index. In 1985, these weights were: food (14.996),
household operation and furnishings (10.6%), transportation (15.8%), health and personal care
(4.0%), recreation, reading and education (8.3%), and tobacco and alcohol (5.5%),
Sources: Statistics Canada, supra, notes 6 and 8.
1987]
NOTES
159
ANNUAL INCOMES AND SELECTIVE EXPENDITURES, 1982a
TABLE 2
(ALL FAMILIES AND UNATTACHED
INDIVIDUALS)
With Mortgage
$30,000-
$34,999
All Inc.
Classes
Without Mortgage
$30,000-
$34,999
All Inc.
Classes
Age of Family Head
Family Size
Total Family Incomeb
Income Taxes
Property Taxes
Mortgage Paymentsc
All Other Expenditures
Inc. Tax as a % of Income
Prop. Tax as a % of Income
Mort. Pay. as a % of Income
All Other Expend. as a % of
38.6
3.6
$32,427
$5,314
$905
$6,160
$20,048
16.4%
2.8%
19.0%
61.8%
40.9
3.5
$38,785
$7,077
$954
$6,493
$24,260
18.2%
2.5%
16.7%
62.6%
48.7
3.4
$32,345
$5,149
$881
$97
$26,218
15.9%
2.7%
0.3%
81.1%
57.2
2.7
$29,067
$4,777
$809
$89
$23,392
16.4%
2.8%
0.3%
80.5%
Incomed
aAll income and expenditure estimates have been rounded to the nearest dollar.
bTotal family income excludes other money receipts (inheritances and lump-sum insurance settlements) and account balancing
difference.
clncludes principal and interest payments.
dEstimate of the “consumption share’.
Source: Statistics Canada, supra. note II.
expenditures in Canada.’ 3 A Class 2-3 Ottawa teacher would have earned
a salary of $33,950 in 1982.14 This would position the individual within
Statistics Canada’s $30,000-$34,999 income grouping. Therefore, excluding
mortgage payments, property taxes, and income taxes,’ 5 the 5.0% allowance
for spatial cost differences would be applied to 62% of the income losses
during the period of house payments (see column 1, last line). Once the
mortgage was repaid, that is, after ten years and until the date of expected
retirement, the cost of living allowance would be applied to an 81% share
of lost income (see column 3, last line).’ 6 To facilitate later comparisons, it
is useful to express these two estimates in terms of a single statistic. The
13The use of dated 1982 statistics to estimate the “percentage” of income devoted to con-
sumption does not appear to be a significant limitation. While the absolute values of income
and expenditure will increase over time, percentage share statistics remain relatively stable.
For a comparison of the expenditure distributions in the two most recent surveys (1978 and
1982), see Statistics Canada, supra, note 11 at 1.3-1.8.
14Ontario, Education Relations Commission, Teacher Salary Grids by Experience and Qual-
ifications, Ottawa Contracts (Printout, 1986).
15See: supra, note 7; below, Part III, Sections 2 and 3.
t6The 62% and 81% estimates also held for the “all income class” grouping, as well as for
the income groupings immediately above and below the $30,000-$34,999 interval; see Table
2. If the damaged individual was a tenant, or if a switch to rental accommodations could be
anticipated, the spatial estimate would be applied to 70% of the annual income losses.
REVUE DE DROIT DE McGILL
[Vol. 33
relocation allowance for price differences on average equals 3.7% of the
annual income losses. 17
2. Shelter/Housing Costs
In the examination of shelter costs, primary attention will be given to
the more difficult case of owned, as opposed to rented, accommodations.
Two factors must be captured by this spatial estimate. First, if a move would
result in higher prices for equivalent housing, then compensation for the
cost difference would appear warranted. Failure to do so would force the
individual to either accept inferior accommodations, or to incur the costs
of an altered, that is, distorted consumption stream through time. In this
latter case, if the housing price difference were to be financed by way of
borrowing, the injured party would experience a series of unplanned re-
payment and interest charges. Alternatively, if the price difference were to
be financed from personal resources, opportunity costs would be imposed;
that is, the plaintiff would be forced to forego a number of pre-accident
opportunities. Such costs may involve reduced consumption of goods and
services, or unintended changes in wealth holdings, for example, a reduction
in income-producing assets and the expected stream of investment income.
The second factor, which stresses the durable or asset qualities of owned
accommodations, suggests that compensation based upon the difference in
housing prices may result in either deficient or excessive awards. For ex-
ample, if the rates of change in housing prices are uniform across the two
locations, and if the award is based upon an equivalent but higher priced
home, then the damaged individual would possess an asset which, when
sold, would yield a higher market value. Relative to the pre-accident situ-
ation (continued ownership of the lower priced home), full price compen-
sation effectively provides a future benefit which otherwise would not be
received.
To incorporate both factors, a present value calculation of the net ad-
ditional costs imposed upon the injured party is required. Specifically, the
shelter estimate involves:
-a current period evaluation of the price difference for equivalent housing in
the two locations,
-an evaluation of the future market values of the two homes (estimates of
the length of tenure within the home and the expected rates of change in
housing prices),
17To obtain this statistic, the 5.0% allowance was prorated by 62% for 10 years (mortgage
period) and by 81% for 16 years (no mortgage period, expected retirement at age 65).
1987]
NOTES
-a present value calculation of the “difference” in future home values (esti-
mates of the discount rate would employ mortgage rates if the cost difference
was financed through borrowing, or interest rates on income-earning assets
if own finances were employed),
-finally,
the net cost (amount of compensation) equals the current period
difference in housing prices less the present value estimate of the difference
in future home values.’ 8
Given the limitations of information in this area, discussion and estimates
are provided on each of these items.
To examine the current period difference in housing costs, use is made
of Royal LePage’s survey of Canadian housing prices.’ 9 This quarterly series
provides expected selling prices on various forms of shelter accommodation
in Canada. Market prices are listed for standardized units (homes possessing
similar features and structural characteristics) and are published by city, and
in the case of larger urban centers, by residential district. The illustrative
estimate is based upon the cost of a detached three bedroom bungalow and
upon a sale/purchase date of April, 1985.20 Under these circumstances, the
individual would experience a locational cost difference of $17,125. The
average Ottawa price equalled $105,375, while the average price for a com-
parable home in Vancouver equalled $122,500.21
The second aspect of the shelter award requires an estimate of the future
values of the two homes, and further, that the difference in these market
evaluations be expressed as a present value. The following information is
used:
‘8This approach also places the individual in an equivalent asset position. The plaintiff
finances any positive differential in the future values of the two homes. The defendant incurs
the cost of any negative differential.
19Royal LePage, Survey of Canadian House Prices (Quarterly). The surveys are undertaken
for over eighty cities throughout Canada. Both selling prices and rental costs are reported on
six forms of shelter accommodation (detached bungalow, standard two-storey, executive de-
tached two-storey, standard condominium apartment, luxury condominium apartment, and
standard townhouse). Statistics Canada and Central Mortgage and Housing Corporation pub-
lish spatial statistics on rates of change in housing prices and shelter costs. However, absolute
price data on standardized units by city are not reported.
20The specific characteristics of this medium-priced bungalow are: three bedrooms, one and
one-half bathrooms, full basement, no recreation room/fireplace or appliances, one-car garage,
1,200 square feet, constructed partially of brick with wood, aluminum or stucco exterior, and
5,500 square foot lot.
21Average prices were calculated across the various residential areas of each city, Ottawa
(East, Kanata, South, and West), and Greater Vancouver (Burnaby, East, Kerrisdale, North,
Richmond, Surrey, Tsawwassen, and West). The average price difference over the four quarters
of 1985 equalled $17,906.
If the precise address of the principal residence were known, the estimate of price differences
would be restricted to comparable homes within similar residential areas of each city.
McGILL LAW JOURNAL
[Vol. 33
-length of home tenure (future evaluation date) of 35.7 years, 22
-annual
rate of housing price increase of 6.3%,23 and
-annual
rate of interest (nominal discount rate) of 8.9%.24
On the basis of these statistics, the present value estimate equals $7,225.25
Therefore, the resulting shelter allowance is $9,900 –
the current price
difference of $17,125 less the present value estimate for the higher priced
Vancouver home of $7,225. Again for comparative purposes, it is useful to
express this spatial estimate in terms of the lost income award. The differ-
ence in shelter costs2 6 equals 1.1% of the annual income losses. 27
22The most straightforward calculation would base the length of home tenure upon the life
expectancy of the injured party. If a shorter time period were to be selected, then an additional
allowance may be required to offset any difference in the costs of rental accommodations. In
the present case, the life expectancy tables for Ontario males were employed. See Statistics
Canada, Life Tables: Canada and Provinces, 1980-1982 (Ottawa: Minister of Supply & Services)
(84-532) at 36.
23Since the Royal LePage data have only been available since 1975, use was made of Statistics
Canada’s housing price/shelter index over the 1965-1985 time period. For both cities, the annual
average rates of change were very similar Ottawa (6.36%) and Vancouver (6.32%). If a com-
parison is made of the two data bases over the 1975-1985 time period, a common rate of
change is present in both series. The respective Royal LePage and Statistics Canada estimates
are: Ottawa (7.96% and 8.06%), Vancouver (7.77% and 7.48%). See Royal LePage, supra, note
19; Statistics Canada, supra, note 6; The Consumer Price Index (Ottawa: Minister of Supply
& Services) (62-001; Monthly); Prices and Price Indexes (Ottawa: Minister of Supply & Services)
(62-002; Monthly).
24Assuming the plaintiff’s own resources are employed, this 1965-1985 estimate is based
upon the annual average rates of return on long-term Government of Canada securities (five
to ten year maturities, Cansim no. B1401 1). See Bank of Canada, Bank of Canada Review
(Monthly).
25Given the nominal rates of interest and housing price increase, this estimate is based upon
a real rate of discount of 2.4%. See Waddams, supra, note 1. Similar to other present value
calculations, the plaintiff’s share of the current period difference in housing costs will increase
with a higher rate of change in housing prices, a lower rate of interest, and a shorter period of
home tenure.
26This estimate does not include any allowances required to offset real estate and legal fees,
the costs of transporting the household to British Columbia, or any initial difference in the
mortgage rates on the home sold versus the home purchased.
27To obtain this statistic, an annuity is constructed such that the annual payments increase
at the same rate as Ottawa teacher salaries (8.0% scale increase over the 26 years of lost income),
a discount rate appropriate for lost wages is applied (8.9% nominal rate), and finally, the present
value of the annuity equals the $9,900 shelter allowance. The 1.1% estimate is then obtained
by expressing the 1985 annuity payment as a percentage of the 1985 income level. An identical
estimate would result if the total shelter allowance is expressed as a percentage of the lump-
sum award for lost income. The data are compiled from the following Statistics Canada series:
Education Statistics (Ottawa: Minister of Supply & Services) (81-002; Monthly); Canadian
Statistical Review (Ottawa: Minister of Supply & Services) (1 1-003E; Monthly). See also Bank
of Canada, supra, note 24; below, Table 4.
1987]
NOTES
3. Taxation, Transfers, and Public Goods
A further source of possible cost of living differences arises from the
public sector. Governments, by way of taxes, transfer payments, and ex-
penditures (provision of public goods), expose individual households to a
broad array of non-market benefits and costs. In Canada, the most com-
prehensive studies of net fiscal incidence have been undertaken by Gillespie.
His most recent analysis, 28 employing base year data for 1969 and extrap-
olations until 1976, examined the distributional impacts of federal, provin-
cial, and municipal government policies. Expressed as a percentage of
income, the incidence of public sector benefits and costs are evaluated across
income groupings and across provinces.
Table 3 presents Gillespie’s estimates of fiscal incidence specific to the
damages example. Following Gillespie, one must emphasize that the tax
and expenditure allocations represent averages across all households within
each income grouping, in this illustration, the $10,000-s14,999 interval. 29
Further, given the inherent difficulties in allocating public sector benefits
and costs, in particular, the use of shifting hypotheses and proxy assignment
variables, 30 any interpretation of individual estimates must be qualified. In
this case, the 1.0% estimate is suggestive of the relative differences in net
fiscal incidence across the two locations (see column 3, last line). Benefits
in Ontario (see column 1, last line) and -8.2%
less costs equalled -9.3%
in British Columbia (see column 2, last line).
Given the extensive data requirements for such studies, as well as their
necessary qualifications, no attempt was made to update the Gillespie anal-
ysis. However, to provide some indication of possible divergencies in gov-
ernment policies over time, comparative levels of income tax (federal and
provincial) and municipal property tax are reported in Tables 4 and 5 re-
spectively. These 1975-1985 estimates for Ottawa and Vancouver are based
upon a common income stream (the earnings of a Class 2-3 Ottawa teacher)
and upon the assumed shelter accommodations (three bedroom detached
bungalows).
Expressed as a percentage of income, these estimates also reflect no
significant differences across the two locations. Through time, both income
and property taxes consistently differed by less than 1.0%.
28W. Gillespie, The Redistribution ofIncome in Canada (Ottawa: The Carleton Library, 1980).
29An Ottawa Class 2-3 teacher would have earned a salary of $12,898 in 1969. See Statistics
Canada, supra, note 27; below, Table 4.
30For a sensitivity analysis of these factors, as well as a discussion of the government neutrality
assumption, see Gillespie, supra, note 28, at 17-25, 60-64, and 117-24; R. Boadway & H.
Kitchen, Canadian Tax Policy, 2d ed. (Toronto: Canadian Tax Foundation, 1984) at 7-15 and
162-68.
REVUE DE DROIT DE McGILL
[Vol. 33
Therefore, in view of the negligible variations present, as well as the
confidence intervals on estimates of fiscal incidence, no allowance is pro-
vided in this case for spatial differences in public sector benefits or costs. 31
4. Summary of Compensating Costs
The consumption, shelter, and fiscal incidence estimates yield a total
allowance in the order of 5.0% of annual income. To offset the locational
cost differences of this example, a similar percentage increase would be
applied to the lump-sum payment for lost income.32 More generally, the
significance of relocation allowances will increase with the value of the
annual income losses, the time period over which the damages extend, and
the degree to which costs vary across locations (relocation to non-
comparable cities).
TABLE 3
FISCAL INCIDENCE, 1969 BASE YEAR,
$10,000-$14,999 INCOME GROUPINGa
Ontario
British
Columbia
Difference
(Ontario Base)
Expenditures & Transfers as a % of Income (Benefits):
13.3%
11.3%
6.4%
31.1%
Federal Government
Provincial Government
Local Government
Total Governmentb
12.6%
10.7%
7.4%
30.7%
Taxes as a % of Income (Costs):
Federal Government
Provincial Government
Local Government
Total Governmentb
21.9%
12.8036
5.1%
39.8%
22.4%
12.5%
4.2%
39.1%
+0.7%
+0.6%
– 1.0%
+0.4%
+0.5%
-0.3%
-0.9%
-0.7%
Net Fiscal Incidence as a % of Income (Benefits less Costs):
Federal Government
Provincial Government
Local Government
Total Governmentb
aDefined on the basis of a “broad income” concept, that is, total income from all sources
-9.1%
– 1.4%
+2.3%
-8.2%
-9.4%
-2.3%
+2.3%
-9.3%
+0.3%
+0.9%
0.0%
+1.1%
before taxes and less transfer payments.
bDue to rounding, the government components may not add to totals.
Source: Gillespie, supra, note 28, at 58, 114, and 141.
31This locational consideration is independent of any tax adjustments arising from changes
in the form and time stream of income, for example, present value differences in the taxes to
be paid on the interest income from the lump-sum award, as opposed to the taxes which would
have been paid on the stream of lost wage income.
32Employing the base data of Table 4 and note 27, the absolute value of the relocation
allowance equals $44,140. This present value calculation does not incorporate any actuarial
allowance for the probabilities of death occurring during the period of lost income.
1987]
Year
NOTES
TABLE 4
INCOME TAXES, 1975-1985a
Total Incomeb Taxes Payablec Taxes Payablec
(Ottawa Teachers)
(Vancouver)
$ Difference % Differenced
(Ottawa Base)
(Ottawa Base)
(Ottawa)
$7,110
$6,630
$6,251
$5,660
$5,188
$4,614
$4,265
$3,938
$3,656
$3,366
$3,356
$0
-$23
-$211
-$232
-$163
-$76
$7,087
$6,419
$6,019
$5,497
$5,112
$4,614
$4,297
$3,998
$3,712
$3,394
$3,356
$39,687
$37,958
$36,368
$33,950
$31,115
$28,518
$26,599
$24,961
$23,430
$21,555
$19,668
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
9a Difference (Annual Average):
2AII income and tax estimates have been rounded to the nearest dollar.
trTotal income refers to the average salary of Group 2-3 Ottawa teachers (maximum grid rates of pay. average of the three
major Ottawa agreements: Boards 52, 129, and 187).
eTaxable income based upon the following deductions: employment expense allowance. Canada Pension Plan payments,
Unemployment Insurance premiums, maximum retirement savings and pension plan contributions, basic personal exemption,
deductions for dependent spouse with two children (minors), and standard medical/charity deduction. Taxes payable include
all surtaxes and tax reductions property/sales tax credits (Ontario) and rebates (British Columbia) are excluded.
dExpressed as a percentage of total income.
Sources: Coopers & Lybrand, Tax Facts and Figures (Annual). Education Relations Commission, Government of Ontario.
supra. note 14. Revenue Canada. Taxation, Taxation Statistics (Annual).
-0.1%
-0.6%
-0.6%
-0.5%
-0.2%
0.0%
+0.1%
+0.2%
+0.2%
+0.1%
0.0%
+$32
+$60
+$56
+$28
-0.1%
$0
TABLE 5
PROPERTY TAXES, 1975-1985 a
Year
Total Incomeb Taxes Payable c Taxes Payablec $ Difference % Differenced
(Ottawa Base)
(Ottawa Base)
(Ottawa Teachers)
(Vancouver)
(Ottawa)
$1,513
$1,441
$1,336
$1,386
$1,430
$1,293
$1,229
$1,166
$1,125
$1,005
$889
$1,401
$1,381
$1,377
$1,381
$1,333
$1,213
$1,161
$1,113
$1,092
$988
$826
$39,687
$37,958
$36,368
$33,950
$31,115
$28,518
$26,599
$24,961
$23,430
$21,555
$19,668
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
% Difference (Annual Average):
aAll income and tax estimates have been rounded to the nearest dollar.
1 ‘Total income refers to the average salary of Group 2-3 Ottawa teachers (maximum grid rates of pay. average of the three
major Ottawa agreements: Boards 52, 129, and 187).
cProperty tax estimates based upon the same shelter accommodations and residential areas specified in Section 2 (three
bedroom detached bungalows in Ottawa and Greater Vancouver).
dExpressed as a percentage of total income.
Sources: Education Relations Commission. Government of Ontario, supra, note 14. Royal LePage, supra. note 19.
-$112
-$60
+$41
-$5
-$97
-$80
-$68
-$53
-$33
-$17
-$63
-0.3%
-0.2%
+0.1%
-0.0%
-0.3%
-0.3%
-0.3%
-0.2%
-0.1%
-0.1%
-0.3%
-0.2%
At least two concluding observations are required on the compensating
cost framework. First, the resulting allowances are “estimates” in the strict
McGILL LAW JOURNAL
[Vol. 33
sense of the term. Given the present state of developments in this area, the
evaluation of estimation errors is, in part, a qualitative exercise. It involves
judgements as to whether the various sources of spatial cost differences have
been adequately identified and included within the analysis, as well as judge-
ments as to whether the employed data bases accurately reflect both the cost
differentials and the circumstances of the damage case under investigation.
The second qualification concerns intangibles, that is, non-monetary vari-
ables which also impact upon an individual’s quality of life and standard
of living. While a change of location would alter the array of social and
physical intangibles, the analysis does not address this spatial consideration.
This exclusion was not only based upon the difficulties of “shadow pricing”
these factors, 33 but also based upon the questionable transitivity of such
estimates. It is not obvious that a priori evaluations are applicable.3 4 House-
holds are unlikely to assign either equal weights, or indeed, similar signs to
a number of intangibles. Consider, for example, individual preferences as
to climate, recreational and cultural facilities, or the proximity of one’s
relatives.
IV. The Compensating Income Approach
The second approach for introducing relocation allowances within dam-
age awards may now be briefly examined. Based upon a compensating in-
come concept, this alternative framework is by comparison very
straightforward. However, in spite of the ease with which estimates may be
obtained, the restrictive market requirements underlying this approach limit
its usefulness within damage awards.
The compensating income framework requires that wages be deter-
mined within competitive and efficient labour markets. If these conditions
are met,35 “other things being equal”, the regional structure of incomes will
reflect locational differences in the cost of living.36 Therefore, if damage
awards were to be based upon the salaries of comparable employees in the
new location, in this case Vancouver teachers with identical experience and
33For a discussion of methods to evaluate intangibles, see R. Boadway & N. Bruce, Welfare
Economics (Oxford: Blackwell, 1984) at 313-15.
34This does not suggest that intangibles comprise an inappropriate head of damages, but
rather that their assessment will depend upon the particular circumstances of the damage case
under consideration.
35Markets are to be free of distortions, employers and employees cannot individually affect
product prices or factor returns, labour and capital are fully mobile, information is complete,
and transactions costs are minimal.
36See B. Fleisher, Labor Economics: Theory and Evidence, 2d ed. (Englewood Cliffs: Prentice-
Hall, 1970) at 205; S. Ostry & M. Zaidi, Labour Economics in Canada, 3d ed. (Toronto:
Macmillan, 1979) at 375.
1987]
NOTES
qualifications, the resulting estimate would automatically include any re-
quired allowance for cost of living differences.
Examination of Labour Canada data provides casual empirical support
for this argument. 37 For office occupations, Vancouver median rates of pay
exceeded their Ottawa counterparts in 32 of the 36 reported classifications.
The average salary differential in 1985 equalled 6.8%. While this estimate
follows the historical pattern of higher Vancouver incomes, 38 it does not
provide a reliable guide for individual salary adjustments. Vancouver/Ot-
tawa differentials are neither stable over time, nor uniform across occu-
pations. 39 Furthermore, if non-office classifications were to be included,
wages would differ by an average of 17.5%.
A more exact application of the compensating income approach does
not overcome these problems. For example, in terms of the present accident
case, where use was made of specific contract information and where pay
grids were matched to the individual’s qualifications and experience, 40 no
significant difference was present in the 1985 salaries of Vancouver and
Ottawa Class 2-3 teachers. Their base incomes equalled $39,330 and $39,687
respectively. This single year comparison does not conform to traditional
relationships. Vancouver rates of pay, as late as 1984, have normally ex-
37The use of Labour Canada statistics represents a “second best” choice of data bases. This
information, while comprehensive, is subject to possible composition problems. The surveys
report October pay scales and apply to all establishments with twenty or more employees.
Therefore, unlike data drawn from collective agreements, the statistics not only reflect locational
differences in rates of pay, but also may capture other intervening variables, for example,
differences in industrial structure or the state of labour negotiations (expired contracts). See
Labour Canada, Wages and Working Conditions in Canada, October 1985 (Ottawa: Minister
of Supply & Services, 1986).
3 8See Frank, supra, note 9 at 5-18.
39For office occupations, a standard deviation of 9.4% applied to the 1985 wage differentials.
40The classification structures for Ontario and British Columbia teachers are not identical.
When recognizing post-secondary education, in this case a four-year degree and an additional
year of teacher training, the Ontario system employs grade thirteen as the base, while British
Columbia uses grade twelve. Ontario pay scales also differentiate between degrees with pass
standing (Class 2), as opposed to an honours designation (Class 3). This distinction is not
present in British Columbia.
The above comparisons are based upon the average pay scales for Ottawa Class 2 and 3
teachers, maximum grids for years of experience (contracts for School Boards 52, 129, and
187), and their equivalent Category 5 teachers in Greater Vancouver (contracts for School
Districts 35-45). The one-year difference in secondary schooling has been ignored. In both
provinces, the next highest pay ranges (Class 4 and Category 6 teachers) apply to individuals
with masters degrees or equivalent qualifications. See above, Table 4; Qualifications Evaluation
Council of Ontario, supra, note 12; Ontario Secondary School Teachers’ Federation, supra,
note 12; Ontario, Education Relations Commission, supra, note 14; British Columbia Teachers’
Federation & British Columbia School Trustees’ Association, Summary and Analysis of 1985-
86 B.C. Teachers’ Salaries and Working Conditions Agreements (1986).
REVUE DE DROIT DE McGILL
[Vol. 33
ceeded the pay scales of Ottawa teachers. 41 The relatively recent erosion of
this differential can be traced, in part, to labour market policies of the British
Columbia government. Since 1982, a series of wage restraint programs have
been applied to public sector employees. 42
The Vancouver/Ottawa comparisons illustrate two basic limitations of
the compensating income framework. First, the interpretation of regional
wage differentials must be qualified. If markets are distorted, if adjustment
lags are present, or if labour and capital mobility are constrained, then the
“real” incomes of comparable labour groupings at particular moments in
time are unlikely to be uniform across locations. Regional wage comparisons
may not solely capture cost of living differences. The peculiarities of local
economies, whether permanent or transitory, also impact upon individual
wage settlements –
differences in resource endowments, industrial struc-
tures, and intangibles, as well as differences in current bargaining environ-
ments (unemployment levels, employer finances/abilities to pay, and
government policies).43 The second qualification concerns estimation re-
quirements. Comparable labour classifications may not exist across loca-
tions. Furthermore, the structure of employee benefits may not be similar
across employers. Given the possible tradeoffs among salary levels, pension
plans, and other non-salary items, estimates which solely examine the wage
component may fail to fully reflect the individual’s employment benefits or
standard of living in the base location. This latter point further complicates
the interpretation of wage differentials. Do they reflect cost of living differ-
ences, variations in regional prosperity (real income differences), or simply
the structure of wage payments as opposed to non-wage benefits?
V. Summary Comment
This paper argues that, in damage cases involving relocation of the
injured party, potential differences in the cost of living must be examined.
If awards for pecuniary losses are to employ a “no accident” paradigm, then
relocation allowances may be required so as to maintain the individual’s
former standard of living in the new location. The significance of such
allowances will increase with the value of the annual losses, the time period
over which the damages extend, and the degree to which costs vary across
locations. Of the two methodologies examined, the compensating cost ap-
41The comparable 1984 salaries are $39,037 (Vancouver) and $37,958 (Ottawa).
42See M. Thompson, “Restraint and Labour Relations: The Case of British Columbia” (1985)
11:2 Can. Pub. Pol. 171. Since 1982, wage scales for Class 2-3 Ottawa teachers have increased
at an annual rate of 5.3%. Vancouver increases equalled 1.5% per year.
43See D. MacAllan, “Should Wage Rates Vary Across Canada?” (1980) 7:1 Can. Bus. Rev.
40; J. Melvin, “Regional Income Disparities in Canada: What They Mean” (presented at
University of Regina, June 1986) [unpublished].
1987]
NOTES
proach was considered to be superior. It not only provides direct estimates
of each spatial factor (consumption expenditures, shelter costs, and taxation,
transfers, and public goods), but also avoids the interpretation and esti-
mation ambiguities of the compensating income framework. This does not
imply that the spatial estimates are free of estimation errors or qualifications.
In addition to the questions raised on data bases, cost sources, and intan-
gibles, spatial estimates are also bounded by uncertainties as to the future.
One cannot reject the possibility that structural changes within the economy
may alter past cost relationships. However, as Cooper-Stephenson and Saun-
ders state:
The difficulty of assessment is magnified not only by the nature of the losses
sustained but also by the period over which those losses may stretch. Thus,
the impossibility of proving those losses with certainty is no bar to the
plaintiff’s claim … . However, it is still incumbent upon the plaintiff to adduce
evidence which will supply a basis on which the probabilities can be estimated.
He must lay the appropriate framework for application of the simple probability
a
standard of proof.”
In this context, the estimates and arguments of the compensating cost ap-
proach should assist in the evaluation of spatial damages.
44Cooper-Stephenson & Saunders, supra, note 4 at 40-41.
